I have had many instances trying out AI where the code it generates is either wrong or a very inefficient way to solve the problem. I've heard from multiple other programmers that AI is mostly useless for them with the software's very complex systems.
@@razorswcyah for self contained or small scripts it's useful but for anything complex you're not gonna get much out of it. If it can be done step by step in small ways and then tied together it's somewhat useful maybe.
Finance guy here: In short, loans had 0% interest in 2020/21 essentially so aka free money for corporations -> corps hire a bunch because they can grow -> money becomes more expensive due to rate increase -> companies can't grow as much -> overstaffed and layoffs occur
@@ansidhe Many of the big tech companies that are doing the most layoffs already voluntarily amortize all their software dev. Section 174 is a big issue for certain types of companies but it isn't driving the layoff trend. The big issue is interest rates.
I’d say also that tech startups never amount to anything. They all go bankrupt. If they’re lucky they’ll bullshit a FAANG into thinking they’re the next hot thing and sell it to them, and 2-3 years later the FAANG finds out the truth and just abandons its development any further. It’s not just low interest rates, it’s that tech startups services and products cannot survive in the marketplace. The jig is up. Venture Capital and investors and FAANGS don’t wanna keep losing money forever.
except it's a specific sector of the economy that is largely seeing this issue, not a general economic recession. Issues with how tech companies are structured and financed (borrowing massive amounts of money, then seeing interest rates increased by the fed) are worth talking about.
You don't have to predict a recession so much as know the hallmarks of a scam. I'm not saying AI/AGI won't replace human capital (given exponential growth, it'll happen sooner rather than later) but shtty automation that is not actually AI/AGI (but that will be called "AI/AGI powered") will first be used to siphon a metric fk ton of money from retail/not in the club yet investors and out of the lower tier economies. There's massive financial/resource hoarding by the wealthy going on atm for highly orchestrated reasons...
@@zeragon7 still utter shit. There's literally thousands of people applying for the same 18 jobs. The market is really bad right now. But legend has it that the jobs may return after September.
@@MarkLitchfield I've had some shit storms lately as well, although unrelated to what you're going through. It's gonna work out. Keep your head up and I'll be thinking about you brother. ❤️
I'm a consultant (so like a contractor who has to pay a cut to the Mob) and four years ago we started this project at my client. We were 12 people including the PM and Scrum Master. We were pumping out features at a good clip for about the first year. Then all of a sudden they started hiring devs left and right and we ended up with over 60 devs, split among 10 teams and productivity just ground to a halt because we're really just working on the same apps and backends, but all in our own little silos. We can't get anything into production anymore because no one has a complete picture of anything that's going on. I was even removed from the business unit that runs this project and put in a specialized unit constructed around this one feature I took ownership of. So now we have our own management structure and release schedule and we have about four standups per day, where about 80% of attendees overlap. It's just a mess.
Your PM team clearly never read The Mythical Man Month by the late Fred Brooks. You can watch a lot of precis/breakdowns of it here on UA-cam. Basic premis: You can't get a baby in one month by getting 9 women pregnant. But he goes into the details of why these teams descend into a mess.
Exactly. Many people here in the comments section are coping by saying the og video is wrong and over-hiring is not a thing. Having too many workers for a project a absolutely hurts the project and those who are disagreeing are either stupid or just coping and making themselves feel better because otherwise they'll have to contend with the reality that there are indeed too many devs in the market, thus reducing the demand of developers.
This is the problem that a lot of companies face - they have big dev teams and nothing is organized properly. The solution that is highlighted in this video is layoffs. You have a team that is too big, layoff right? But what if we didn't? What if instead we reoganized to work on bigger and better features. What if as developers companies hired juniors to collect data to improve and refine a product. Too many developers? Hire some social media managers or marking and business assitants to help with the end product. Devs too isolated in their own bubble? Get some PMs to work with these devs to bring them back to the full picture. Devs today are tasked with a great deal, so what if we got better about supporting the company as a hole by hiring more support roles?
"a contractor who has to pay a cut to the Mob" is probably the best way to describe working as a consultant😆 I used to work as a consultant and I wished that project managers could be replaced by A.I, those mfs consistently made highly illogical decisions
7 years ago a professor at a notable university explained that new age business practice is firing all the people who are paid the most (and the most experienced). This is so they can hire new people for less than half the salary. He explained that it's leading to worse quality of everything. Now here we are with a growing demand for quality products. :)
I was laid off and the statement was it was because of my salary. Please take my advice and work/save to retire as early as you can. It is so nice to reach a point where it doesn't matter what happens. I wish I started saving earlier, I could have retired 10 years ago.
People wonder about why every company lays off 10%, and it's so #1 workers are scared #2 job market is weakened #3 new employees are easier to indoctrinate
AI powered scam calls are the best use case right now - you can get a very human sounding AI powered chatbot to call people and ask for donations, while pretending to be a nonprofit etc
@@PowerWinsTop This is absolutely the risk of AI. It is very good at making things that look and sound right but is entirely fictitious. AI creates false information and scams generated at a way higher rate and far less recognizable than we've had before.
In terms of capabilities yes, but cost not really. Those workers are getting fuck all in pay, meanwhile ChatGPT's API pricing is extraordinarily high and just a single GPU to do it yourself (at reasonable speed, let alone real time) is 5 digits. Then include the fact that you also need to generate a voice for that text (TTS would get hung up on instantly) and have the hardware and infrastructure to handle all that audio data going from your servers to the phones. For the scale these call centers work at I think there'd need to be more developments in terms of speed/price over quality
I read somewhere that the big tech companies hoarded talent as an anti-competitive move. The layoffs are a reversal of that talent hoarding move. This creates an opportunity for start ups.
People assume companies make smart decisions. They don't always. They saw Twitter layoffs and executives were like oh shit maybe we don't need all those people. In a couple years when all those companies are getting hacked and having failed deployments they'll be like oh shit that's why we hired all those people.
Nope, after a project is finished and has entered maintenance mode, you only need a fraction of developers to support it. The base product Twitter has already been built, and will only need a fraction of devs to support it now. So as rude as it is for Elon to fire them, Twitter really doesn't need that many devs anymore.
@@Dipj01 also, as rude as it for Elon to alienate users - less users means less development and maintenance. Also as rude as it is to alienate all clients paying for the ads, less money means less need to count the money.
The thing about "coding AI" has always been funny to me because you need engineers to understand your tech stack to actually accomplish the goals of a given business need or strategy-something that necessitates human to human understanding and interpretation. It will be an assistant for quite a long time, at least for anyone doing backend and full stack.
If a programmer is replaced with the current state of AI, it's because, either that person should have been fired regardless of AI, or the person who fired them should have been the fired one.
Yeah watched the whole video and don't think bro said anything profound or even accurate. I actually even reject his conclusion that you should "become more generalist to make the company valuable". How far is having surface level understanding of 10 frameworks going to get you? And if the answer is "well, don't just have a surface level understanding", then I implore you, how do you bend space and time to gain any notable experience in everything? I see so many of these tech commentary videos nowadays and at the end of all of them, I'm always left with the same question of whether or not they even work in the industry or whether or not they are grifting.
@@poopymcfartbeanthat's probably a bit of hyperbole, but I agree with his take. The software companies have a hell of a lot more developers than what they produce. Having huge number of developers beyond a certain point actually hurts the project instead of helping. The classic adage of too many cooks spoil the broth still holds up.
as a math PhD student AI is not good at math in the same way a human can be good at math. it's very good at doing the kind of math you do in a calculator, but doing pure math like topography (my specialization) it just isn't all that capable of doing. but I guess this is the difference between a mathmatician's conceptualization of what math is and what everyone else thinks is math.
@@9s-l-s9 I think you're confusing Prime saying "don't be just a frameworker" vs. "don't be just a domain specialist". Prime is a domain specialist at Netflix.
It is true. The risk is that you become dependent on your niche. If it shrinks or collapses or goes obsolete - you might have a difficult time finding a new job.
That guy used "specialist" in the wrong way. A specialist is someone that excels at and is especially knowledgable in a specific area. But he doesnt talk about specialists, he is talking about hyper siloed roles.
@colincotterell3365 great UI people can actually help form strong requirements and even save companies a ton of time and money by putting up UX guardrails to minimize user error and also by reducing risk of getting lawsuits based on usability
my hypothesis is that corps got a bunch of free money during covid and expanded gangbuster and then once it went away they realized they had too many engineers and axed a bunch
Govt spending was out of control, along with foreign policy, in turn the speculation that justifies tech hiring booms is less reliable. The EXTENT of the AI craze is a gamble, and confidence of meeting necessary EXTENT is quietly lacking.
@7:28 You nailed it. I'm a computational physicist and control systems software engineer: trust me when I say ChatGPT is shit at math too. I think most people just don't dive deep enough into math with ChatGTP to see the hilarious gaffes. Moreover, if I were to “accidentally” let those gaffes into my code, people would get hurt. We don't generally use things like ANNs to compute control signals in safety critical systems because we need analytical techniques to prove certain theings about program behaviour (not that you could even get insurance for it, even if you found an engineer who would sign off on it). So letting AI wirte the software in the first place is a LOOOOONG way off.
The people that use AI for code probably aren't good at auditing, or even know what that is. AI code should be put through the same filters as any other foreign code submission (like all open source code contributions). This sadly defeats the point of using AI to begin with, that is for the majority that just wanted to do no work and still get paid.
Does it even need a deep dive? Tried asking it about energy needed to boil some water in a microwave. It gave a different answer with each click of regenerate despite seeming to use the same formulas each time. I've only tried using it for things that I have at least some rudimentary knowledge about though. If I asked it about some legal stuff I would not be able to spot any issue ...
Was helping a friend with a pharmacodynamics problem not too long ago. Chatgpt couldn't solve a pretty easy to calculate formula ln(C_max/C_min)/delta_t. Got it completely wrong every time. I even gave it the correct answer. Different answers every time. It at least got the general idea correct. But...yeah. For more esoteric things that don't have a whole lot of training data to develop the model, I highly doubt it'll be any good. That is unless true, generalized AI actually works where it literally thinks and functions like a human with a 160 IQ, or more. Highly doubt that'll ever be a thing though.
What Thor's been saying specifically regarding Godot vs. Unity is that while "Unity - the Engine" is still kickass, "Unity - the company" has in very recent times made decisions that make them untrustworthy. Even if they retracted their utterly insane monetization plan, we can't trust them to not make stupid decisions again. Decisions that might mean they're not gonna be around in x amount of years.
+1 for encouraging kids to do highly complex manual trades. My father was an engineer and welder. He could draw stuff and then weld it together. Fuck my "IT skills".
I feel like the guy in the video is a good example of fake it until you make it. Super confident, very professional video, and yet has no idea what he's talking about.
button count proportion makes sense. There should be at most 1:2 (button:engineers) so that at no point in time a button is without an engineer. No button left behind.
Yeah his point about being a generalist makes little sense to me. Whats the point of working on 10 different frameworks? At some point, the requirements of the company will demand specialization in different areas so your surface level understanding of things will no longer fly.
@@fuckmyego sorry man. I just lucked into the job. I've got two pieces of advice. 1) never roll your own security. Always use open sourced vetted libraries to do it. 2) by never I mean never. Unless you have a PhD or equivalent work experience
The stock market works on forward projection, not current value. Therefore, hiring more devs can significantly increase your company's "value" by projecting strong forward growth and product release. At some point, you have to either deliver or make major cuts before the bubble bursts. The latter solution seems to be tech companies preferred. By doing this the company can have overvalued forward projections, and if they cut in time, they will get a second boost in evaluation for cutting costs. The workers are mostly pawns in this money-printing scheme.
The point is not that Facebook did this intentionally. They goofed. Most Valley companies goofed. They did not correctly anticipate growth decline because it never happened to them before. The valuation increases is because they actually turned around and fixed it instead of bleeding and having way too low EPS and gross margin.
Yes, it‘s a cycle, because investors are shizofrenic. When the fundamentals aren‘t there, and you don‘t have staff to fire, you hire a bunch of people, projecting future growth. If that doesn‘t work out, you now have a bunch of people to fire, projecting future growth. Rinse and repeat. This way a company, mainly in tech, can make the line go up for a long time, while just treading water on the fundamentals.
@@dv_xl they only didn't antecipate the decline if they were to dumb to understand how that growth happened, or were they counting on people working from home forever ? because the same companies dont even want their own workers working from home anymore
Wait...they dont talk about this in the video? This is the main reason for the layoffs so the fact that they didn't even cover it makes everything else theyve said suspect...
Judging by these responses, it's clear people are confused about the cause and effect of certain events. Stock goes up because layoffs does not mean layoffs occur to make stock go up. If that was the case, take it to the logical extreme, no companies would have any employees.
Do some game dev myself and needed to calculate the physics of an arch projectile driving an arrow with set height and ai failed, finally found a physics lecture in UA-cam and was able to derive the correct equation, still has ways to go
I tried to get it to help me write an implementation of hookes law for a vehicle suspension and it reversed the sign so that spring pressure was max at max extension rather than contraction. It took me a couple hours to figure out what was wrong as I'm pretty new at gamedev and the math around it (this is a game to learn this all on). I eventually just read the wiki page, understood it, and implemented it in less time than it originally took with GPT. Additionally, because I spent the time to learn it myself, I now have new knowledge on 3D math that I have already applied in other features. Thanks for wasting my time GPT.
Thanks for putting out this video. When Josh put that video out, I was shocked at the amount of bad information. It doesn't take "shockingly small number of software engineers" to build and maintain the modern applications like Facebook. I'm glad someone with a UA-cam following made a counter video
People that say that it is an easy work to make a button/form/etc. - haven't really made one, right? Making a fully-featured set of controls that needed to make an excellent form - it is hard. And i'm not even talking about the fact, that making uikit is not what most frontend devs do daily. Localization, accessibility, internal tools...
@@jimbeam9504just because you managed to land a job doesn't mean a job crisis doesn't exist. People not landing jobs doesn't mean they're not "good". Survivorship bias at play there, mate.
31:15 Hard agree with Prime. Honestly, even from the perspective of a Gamedev, this is still shortsighted. Most games won't need 100 people, but the truth is, there isn't really a limit to the feature-set of a game. Even taking a simple game like Minecraft, you have: -World generation -Interactive blocks -Movement with / without items, while / while not riding, all of the movement stuff -Crafting, enchantments, repair, etc -Fighting (armor, potions, effects linked to enchantements) -NPCs -Shaders & VFX -SFX -Optimization -Network code (LAN & server) Oh, did we forget the dimensions? And the command blocks? And the events? What about the tutorial? Of course the recipes. Wait, there's a non-java version? And console versions? And even a phone version? None of these tasks are hard or very lengthy, but they can be worked on by different people at the same time, which is the most basic requirement for hiring employees. So turns out, if you want to make a Minecraft clone really quickly, you could have maybe 20 or so engineers. No one would hire so many people to deal with so few simple tasks, but it's possible. Now how would you even manage to convince anyone that a project can't be split into more than 100 parts?
he is pretty naive because he is a game developer, I was a game dev then went to web for a better life, but it is incredible the difference between people (humble people) compared to the no soul that exists in the game developer industry. The whole mentality is that everyone dont know sht since they are not doing game dev.
Based on your channel, you are the farthest thing possible from calling your self a game dev..... script kiddie is a more apt description of your talents.
On feature delivery times: localization itself is a feature that greatly impacts the delivery time of next features. Last project i worked on implemented it and suddenly we had a new step before release - send the feature to the localization team. It added between 2 and 5 days depending on what we were delivering
6:35 I think his point was more about the creativity argument, and I fully agree with it. People love to think of creativity as something unique that AI could never replicate, but this is severely mistaken. Precision is indeed the hardest problem that likely won't ever be fully solved, as it has infinite definition, but what we call creativity is really only the sum of our experiences mixed with randomness. AI is currently not great at being creative, but it is a finite problem. Maximum creativity is random noise, and what we consider as being creative is within bounds. There's the randomness you want (creativity) and the randomness you don't want (bad creativity), being creative is just having an sense and intuition for a subset of the valid creative space.
Elon hate is media brainwashing. Dude is literally the most successful entrepreneur of all time. Nobody else has started an American auto company in recent history. Few have created space companies (and SpaceX is on top). Starlink is bringing the Internet to everybody, everywhere. Elon buying Twitter allows free speech on a platform that previously was caught suppressing "unpopular" views in collaboration with CIA (see twitter files). Everything I've stated is known facts, and all you've got is "lol 40 billion big number"
Chat Gippity is a tool, like anything else. Yes, you need 100% working code (which AI rarely provides), but it doesn't matter. The point of it (and CoPilot) is to write some boilerplate code, as a "springboard" for you, in order to reduce dev time. Also, Chat Gippity is great for getting answers to high-level concepts. It's not just good for writing code, but I'd say it's almost spot on as a learning tool to give you context. DISCLAIMER: Yes, don't trust it. AI hallucinates. Always vet and test its answers. Just don't throw the proverbial baby out with the bath water
Brett Victor gave a brilliant presentation called "the future of programming". In it he talks about coding in goals and constraints, not instructions the computer should figure those things out. This concept is over 50 years old now, maybe AI is how it will be unlocked. So not replacing programmers but changing the way they work
51:00 a previous company for who I worked had a c level employee 'reply all' everytime he needed his assistance betty to post on social media. he was also the reason why the implementation of 'password change every year' because he kept clicking on phishing mails....
Thought experiment...let's say we didn't have any type of sorting algorithms yet, but we do have AI. Could AI come up with all of them? If yes, why hasn't it come up with any new ones yet?
Well, its a bit more than a complex approximation function. First, its a recursively self calibrating function. Then after a while you realize that its not only approximation, its actually a deep linear regression, that ultimately will linearize an input set of non-linear values into recognizable output of values of a well defined manifold.
I was half way through the wheel of time books and i asked the AI about my progress because i had some questions, and it was telling me all kinds of crazy shit that never came true.
i got laid off jun 5. the silly part is it had nothing to with the trends in the industry. our ceo was caught embezzling money and we lost all of our contracts. now im dealing with the state of the rest of the industry. life is funny that way. pray for me
I absolutely think *some* programmers will lose their jobs to AI. Especially in non-tech spaces. IT/Engineering HR competence in non-tech environments is not great. I mean if you're on COBOL and IMS you're probably safe from AI, but if you're specifically doing reasonably modern tech at a place run by AI enthusiasts there's probably a non-zero risk. However, that's pretty niche and even then I doubt the threat is substantial.
I have a pdf copy of a paper written by Nathan Papke that talks about AI with great detail. It even includes some C for loops that mimic rudimentary neurons. I didn't know it was so long running until I read that.
A lot of tech companies have workers who legit do nothing. (If you don't think this, you are one of the people who does nothing) The higher-ups are slow to realize that their company has bloated, seems to come in waves.
The thing is it will not flat out replace you. It will make your job take less time. Which probably means less programmers will be needed. Hopefully this is offset by all the jobs AI created for when building AI-stuff. But it's best for the very tedious stuff. Like populating a large json or struct.
Like in many earlier instances of productivity increases, i'm not even certain the total amount of programmers will necessarily reduce, atleast not for a while. We'll just increase the scope of the things that we're trying to do instead.
It boils down to a single reason: money. Layoffs occur because companies can no longer write off 100% of software developers' salaries against taxes. Have a look at Section 174 congress act.
7:00 there's a whole lot of people making decisions in business that are trying to get by with "close enough that others don't catch on," and the truth is that it catches up with you, instead of people catching onto you
On whether AI can solve new problems: we're used to thinking about AI as a tool to make creative works from prompts, but a lot of the power will be coupling it with skilled humans in a conversation. Instead of just Jeff Dean and Sanjay Ghemawat pair programming together, imagine them programming with Gemini. Or imagine Donald Knuth paired with an AI. Or check out Terence Tao's explorations with using ChatGPT in proving theorems. Likewise, while generated code is currently a nightmare, I suspect people may underestimate the savings we'll get from intermediate and advanced engineers using AI in the design (rather than coding) phase.
Wow somebody doesn't understand how AI works... it isn't thinking... it is a probability machine that based its probability on a bunch of stuff that people wrote. The only thing that is being underestimated here is how dense you are.
I have spent time pair programming with Gemini. For very small snipets in simple languages like Python, it is okay at best. For anything else, it is terrible; Calling methods that don't exist, using types that aren't defined, making copies of heap allocations where they are not needed, exponential time solutions to linear time problems, testing things that don't need tested, not testing things that do, using third party libraries that don't exist. Design isn't much better. AI works great for 'best fit' type of problems. Programming and design don't generally fit into this category.
@@ants_are_everywhere What makes you think it is? Yes probabilities are taken into account but you don't make sentences by choosing the next word you are going to say by taking a list of probable words with percentages attached to them and then taking a random number and choosing that word because the number told you to and it might have been the least probable and therefore nonsense. Stop it with the bull.
The main reason for the Tech layoffs is the increasing cost of capital due to rising interest rates. For the longest time, capital was free and making profits could be delayed until the future. Now capital costs money every day, increasing companies incentive to shift focus from stock market value to actual earnings.
AI developer and autonomous aerobotics engineer here. Def not in an nft FOMO type deal due to actual physical MVPs in our world. There will still be a need for software devs, engineers and programmers just like we need car mechanics for our cars.
I totally get with the 'NFT era of AI' take. GPUs getting pricier by the day, companies funded by investor hype and not much else, it's echoing that era so perfectly you'd think it's parody
@@andrzejostrowski5579 i mean cloud computing was very much a game changer, you will be hard pressed to find a startup today that doesn't run in the cloud. NFTs on the other hand...
I've long held the belief that corporations tend to over-hire during economic booms, possibly with the intention of identifying and retaining the best employees while also providing a buffer for future needs. However, when economic conditions shift, they may seize the opportunity to part ways with underperforming staff under the guise of economic necessity.
To answer the question about why you still need a software dev for things that already exists. It's because a normal person doesn't think like a developer. A normal person can't even put in the right prompts to get the desired output. They need to think about all the small things, how it flows, common features, and missing features. AI can't really code that well right now, but even if it could, people can't think like a programmer. It's like having a team full of bright people, but they have an ignorant lead who doesn't know how to use their skills.
There's a problem of personal interest No manager says "hey boss, I have too many engineers, you can take one of my best and move her to another team that needs talent" No director ever says "Hey boss, I have too many teams in my org, you can take one team and disperse them to other orgs" Managers always want more money more engineers more resources, and inevitably they will overhire if the company lets them
42:00 is probably the best advice you can give new or Jr programmers from my experience. you can get really good at something and that's great, but being flexible and having the ability to understand what people are saying across the board saved my ass more than anything else.
7:16 Physics major here, AI is horrible in math. Not only can it not do any advanced stuff like Lagrangian mechanics for even the easiest problems, but the other day it couldn't generate a system consisting of 3 equations with 3 unknowns (essentially a 3 by 3 Ax=b problem) which I asked for my 9th grader student.
Tech companies are laying people off for a few reasons: 1. Investors love it as it increases short-term revenue, so tech companies will continue to do it while they're being rewarded. 2. Employees are costing more due to wage inflation. 3. Related to #1 but Interest rates are higher which is hitting the bottom line meaning laying off is an easy way to show you're focused on maintaining your margins. 4. None of the big players are investing in creating new traditional apps or services, they're focusing on growing what they already have. You don't need as many engineers for maintenance as you do for creating something new. 5. No one is creating traditional apps + services because investment has swung to AI, no one wants to invest in the "old" stuff anymore. Most apps that involve AI don't need that many engineers to make as they're all utilizing existing compute architecture (which handles large-scale distributed inference and has existed in AWS, Azure, GCP, etc for a while) and most AI apps are just relatively thin wrappers around calls to LLMs.
Very interesting takes about generalizing and specializing, I've pretty much always heard that it's better to specialize in something because you'll be incredible at that one thing and in the same way be hard to replace
There is such a thing as specialization death. If you are not generalized enough if the market changes and what you are specialized in is no longer needed you are in a dead in. Both over qualified and under skilled for available jobs.
@@thomasmatthews7388Of course but that just means you need to be able to adapt to the market demand. I don't think the solution to this is to be a "generalist" because you will only ever have such limited understanding of the things you work on. Unless the company isnt evolving, eventually the requirements of the company will require you to start specializing in certain aspects and the generalist approach won't cut it. Specialize, but don't put all your eggs in one basket is the better approach.
@@thomasmatthews7388 But that implies that you've already hit the threshold of specialization, and then you start diversifying. Not that you're the guy who knows a little bit of React, a little of Django, a little of SQL, a little of devops, and you're trying to be a one-man band pretending that you can ship at the speed of 5 people.
More than 1 hour, but a valuable content!!! The layoffs are more economic than AI or other phenom. Learning fast is a huge benefit everyone that is a beast that I knew they have these particular skills, they're learning stupid fast new things!!
He doesn't have the full picture on the layoff, which he says are bad for the company, but there's an investor hype cycle these tech companies leverage to do enormous buybacks right after layoffs.
Questions I have about this: (that I will be looking into anyway but anyone please let me know if you have the answer.) 1. Are these 1000's of tech layoffs all definitely developers/engineers? 2. What does the job market look like after this, are people fighting over a small job pool or will we see a lot of new start-ups trying to disrupt big-tech? 3. Is freelancing a better path for the industry as product requirements seem to vary too much to hire permanent staff?
I recently said something stupid like "CEOs do nothing. They spend 80% of their day in meetings. I spent a lot of time in meetings too, and nobody's paying me a million dollars a year." I got a good reply: "the fact you don't see the value in what a CEO is doing is the reason you're not being paid a million dollars a year".
I'm not an expert when it comes to the economy or Wall Street, so is there anything that prevents a company from "burying" their real revenue/profit margins with the cost of hiring a bunch of people for the sake of "preparing for a future project"? I was under the assumption that home investors/professional investors could be tricked into thinking a company is more valuable with employees, not just because of the increase in headcount, But because the cost of hiring those employees might be used to fudge other numbers. I'm pretty sure that I'm wrong but I just want some help understanding why.
It's hard to hide the lack of growth. When companies are burning cash ( e.g. Amazon in its first decade or so ) they are given quite a bit of leeway if they are demonstrating significant growth. The concept being that funding growth is costly ( new regions, new product offerings, marketing / customer acquisition costs, etc ) and if the growth is there costs are discounted. Although if costs YoY are outpacing the growth metric, then the knives come out. So no, you generally can't hide a fundamentally bad business by increasing costs uncorrelated to a growth metric that the street accepts. Now this doesn't mean that some bespoke growth metric isn't B.S. , but the street's analyst community usually smells the 💩 pretty quickly.
They can afford to fire X% of their workforce even if they have temporary losses, make whoever stay work more hours or more intensively and, in the next decade, tech salary will decrease for everyone, as there will be a bigger mass of unemployed developers desperate to work for much less than what they used to. Every industry has done that over the past decades. It's just for greedy profits.
and yet there is a programmer shortage :/ I think skill comes into it, lots of new frontend devs, not as many backend with the experience companies need. the pay will likely be polar opposites for a while
there is always a shortage of everything if you ask companies. if i may translate that for you, it means: "we dont wanna pay our workers, please overflow the market with supply" @@Mel-mu8ox
Didn't even know TechLead said that. I watched a few videos like 4-5 years ago, found that he centers everything about him having supposedly been at Google and now a millionaire, found him annoying mainly due to his personality - and clicked "do not recommend channel" every time I've seen him again. Good to see that I had an accurate reading 😂
Right now it’s mostly devs not being hired rather than devs being layed off. Thats because companies are anticipating that AI will be much more intelligent in the near future, so there’s no need to hire right now. Eventually it will replace devs in a few years as it gets more intelligent exponentially. I know he doesn’t agree but it’ll happen just wait and see
1:11:15 While both Nokia and Ericsson left the consumer electronics market, they are still the second and third biggest vendors of telecom network infrastructure at around 13-15% market share each. The pair are building most of the 5G infastructure in countries like the US that banned Huawei, and Nokia Bell Labs is even building a 4G network for NASA on the moon.
I think we should stop considering head counts in tech space. It just doesn't make sense. One person created Instagram but his motivation was on a different level and now you need 100 employees to mange the same app because not all devs have the same level of motivation. They just work 9-5. And there is nothing wrong with that.
And AI may also soon replace your job. These dummies who use today's AI as reference and assume nothing will progress non-linearly are batshit, utterly.
I belive that a major reason for layoffs in Tech is simply because they over hired people during the tech boom a few years ago. And now there are so many useless positions occupied by people with little or no reason to be in the company to begin with. DEI "officers" at my workplace for example, where all fired because they never did anything that would profit the company. People who go to meetings just to be in meetings also got fired, they never actually provided anything of value to the company. This is also why Elon fired 80% of Twitter when he bought it, and it still runs just fine, because the majority of people working there weren't necessary for the company to work.
Did they rehire new DEI officers? Hopefully they did but if not, thats very concerning. DEI in this industry is the worst I’ve seen/experienced in my work life so far. I haven’t worked in too many industries but its a major problem.
@@xmissyangelzx They have not rehired the DEI officers. But as a response we have apperently lost some funding from the orginazations that demanded them hired in the first place. But overall its still a money saving decision. The DEI officers did nothing but hinder the hire of great employees because they didnt fit the quota. Being Diverse (Not white, straight and male) is what they strive to hire. This is terrible for buissnes in general, because most people in gaming are straight white guys who do nothing but game all day, and to refuse to hire those with the passion to work for games or codes because they are not meeting the DEI requirements are a loss to the company. The NGO's who wanted us to hire these officers in the first place has made sure we will never be in the Game Awards, or at the very least threatened to.
@@paragonaesir1957 I understand that and it is a fact that most developers are white men/asian men (unsure about sexuality as thats harder to track). But they shouldn’t have too much trouble finding people that arent white /asian men for other roles in a gaming company. Im not in the game industry, Im just a mobile developer. But I will say DEI is really important for the gaming industry because gamers come in a variety of people that aren’t just white men. When you have a diverse team, it brings new perspectives which lead to new ideas, as well as accurately represent the global audience that makes up gamers. I understand the company you work for doesn’t value it but thought to say this. DEI isn’t just about hiring diverse candidates, it’s also about training people appropriately to know how to treat others different to them. Discrimination is such a personal issue of mine when it comes to my work, Im at the point where Im gonna leave the industry if I can’t land a job in FAANG because these are only set of companies I see that actually put in effort into DEI, particularly on the engineering end of things. I hate the Data structure and algorithms stuff i have to learn but I see no point in me continuing this job, if I have to feel like Im fighting with my coworkers all the time, just to be taken seriously at my job or even just simply do it lol. When I mean its such an issue, I notice these issues from interviews I do as well as every place I’ve worked in this job at so far. Im very much an outlier in this industry. But its been more of a handicap than anything. In comparison to my white male work colleagues they don’t have such hard time with how people treat them, despite us all being as competent. Also in general, most women Ive worked with have mentioned instances of not being treated well, for unjust reasons. Theres a-lot of sexual harassment, bullying, racism, sexism, (i dont see much ableism as im certain loads of devs are autistic😂) but theres just alot of bad behaviour people do intentionally/unintentionally that people get away with because of the lack of effort put into DEI. Theres also statistics that most women leave the industry. Anyway I don’t mean to ramble. Im not into the idea of hiring someone to fit a quota but most places won’t hire someone that cant do their job as businesses need money lol. But I just thought to put it out there. I can’t speak for how your workplace handled DEI as Im not there but thought to just give you another perspective.
So, ~20 years ago when I people used to ask language trivia in interviews to show mastery. When I did my most recent round of interviews at FAANG companies, they didn't even care if I was familiar at all with the domain or language. They just wanted to test for general coding ability and problem solving. So right off the bat this guy is not even a little wrong, he's Uncle Bob levels of wrong. He then says "social media sites used to be developed by 1 person"... I actually think this guy is a master troll.
Wait till he gets to the part where he says if you have over 100 software engineers "you're done". As if 100 is optimal, but 101 is not. He is probably a troll.
AI doesn't replace programming jobs; employers do. They could be socially responsible and retrain people instead of firing people and bring on new people who aren't familiar with the company and require costly on-boarding.
My concern with AI tools, is the same I had with outsourcing jobs. The advanced jobs aren't going away, but advanced programmers all start as entry level programmers who get better over time. If you need to have advanced skills in order to get a job, a lot people aren't going to get the real world experience necessary to get better and a lot of people won't even get into the industry at all which will lead to longterm problems finding advanced programmers.
AI has shown me the arrogance in the programming community. 5 years ago people were saying artists were gonna be the last to be replaced by AI because their work requires creativity.
I know companies, not yet public, that use VC funds and hiring people like crazy. And they do all this show to portray a big company. So, Josh is right! Bang on target
7:15 I decided to test this so I paused the video and threw it a few algebra 1 questions and it surprisingly got them all right. (I haven't used chatgpt in a while but I remember being terrible at math) so after that I thought to myself about how chatgpt was better with polynomials than me, unpaused the video, and was told that I sucked at math lol
I put in it a high school problem which had calculus 1 and differential equations.It could solve the differential equation because it was classic and there were online for similar types. But when it came thinking outside the box it failed. You just have to throw it problems made by a mathematician in your region(Each region has different philosophies of math problems) and will fail miserably. Don't just use USA derived math questions.
@nikos4677 Yeah, it more or less just memorizes the answers and looks for a few basic patterns. It doesn't really understand what it's trying to solve or what for. But it still did surprisingly well, in my opinion.
100 programmers is not a stupid take. Look at natural human social group sizes, it matches a lot of things, you have 4 good friends, 10 okay friends, 25 friends of friends you can hang out with, 40 people with whom you might have to interact this week, 100 people you have in your phonebook for this month and 200 to 400 people is all you know. These are all natural unit sizes for humans used in the military, company organization and school organization. Wherever you go if a school has more than 50 people that's two classes of 25 people and all of them will be in friends groups of 3 to 9. Those are just platoons and squads or departments and task groups. If your company has 100 programmers and 4 to 6 departments with 16 different task teams, bro that's the most you can go with one CEO naturally. Everything above that the CEO relegates remembering names to managers. Up to 100 he can remember all the names actually.
Going deep into the rabbit hole here, Prime. War is costly, and cost cuts always come from where the rope is thinner: the working class. This has been so for millenia, and if someone is interested: "The Theory and Practice of Oligarchical Collectivism", the fictional Book wrote by Emmanuel Goldstein, the public enemy #1 of the Party in Goerge Orwell's "1984", explains the role wars play in contemporary capitalism quite well.
GH Copilot hallucinated a configuration entry that looked "right" and went to prod, I've spent the entire day debugging it. It created more work!
Good point. I also see that coming, debugging skills will be even more demanded.
I have had many instances trying out AI where the code it generates is either wrong or a very inefficient way to solve the problem. I've heard from multiple other programmers that AI is mostly useless for them with the software's very complex systems.
honestly about the only thing I have seen it be useful for is generating the boilerplate/overall structure of code
@@legokill1019 For me chatgpt made up a nonexisting plugin as a problem solution ...
@@razorswcyah for self contained or small scripts it's useful but for anything complex you're not gonna get much out of it. If it can be done step by step in small ways and then tied together it's somewhat useful maybe.
Finance guy here: In short, loans had 0% interest in 2020/21 essentially so aka free money for corporations -> corps hire a bunch because they can grow -> money becomes more expensive due to rate increase -> companies can't grow as much -> overstaffed and layoffs occur
This, plus Section 174.
@@ansidhe Many of the big tech companies that are doing the most layoffs already voluntarily amortize all their software dev. Section 174 is a big issue for certain types of companies but it isn't driving the layoff trend. The big issue is interest rates.
I’d say also that tech startups never amount to anything. They all go bankrupt. If they’re lucky they’ll bullshit a FAANG into thinking they’re the next hot thing and sell it to them, and 2-3 years later the FAANG finds out the truth and just abandons its development any further. It’s not just low interest rates, it’s that tech startups services and products cannot survive in the marketplace. The jig is up. Venture Capital and investors and FAANGS don’t wanna keep losing money forever.
@@fennecbesixdouze1794 we can interpret it as this. value abstraction has more abstract value to get temporarily an abstract value
this was written by ohio gang btw. me and martin. it goes hard
Predicting recessions is like predicting that you're gonna die someday.
Yea but when you go to the doctor and they tell you, you have cancer...
not about predicting recession, there will be one for SURE, we just dont know the when. Our econoy goes in cycles, ups and downs, thats for sure...
except it's a specific sector of the economy that is largely seeing this issue, not a general economic recession. Issues with how tech companies are structured and financed (borrowing massive amounts of money, then seeing interest rates increased by the fed) are worth talking about.
You don't have to predict a recession so much as know the hallmarks of a scam. I'm not saying AI/AGI won't replace human capital (given exponential growth, it'll happen sooner rather than later) but shtty automation that is not actually AI/AGI (but that will be called "AI/AGI powered") will first be used to siphon a metric fk ton of money from retail/not in the club yet investors and out of the lower tier economies. There's massive financial/resource hoarding by the wealthy going on atm for highly orchestrated reasons...
Economists are split between the optimists who think recessions are a thing of the past, and pessimists who predicted 10 of the last 2 recessions.
i love that his hair acts as a half-green screen lol
Typical cyyber
That's the true transparency many influencers cannot achieve. Prime literally let's us see through his brain
@@Thectbut there's nothing there 😢
Teals you something about the person!
He did it on purpose
This was so relevant today. CEO announced layoffs this morning and I got my invite with HR shortly after.
Stay strong brother.
That sucks, most of us have been there. Best of luck finding a new place, hopefully one that’s stable.
This was 5 months ago. Are you okay? How are things now?
@@zeragon7 still utter shit. There's literally thousands of people applying for the same 18 jobs. The market is really bad right now. But legend has it that the jobs may return after September.
@@MarkLitchfield I've had some shit storms lately as well, although unrelated to what you're going through. It's gonna work out. Keep your head up and I'll be thinking about you brother. ❤️
I'm a consultant (so like a contractor who has to pay a cut to the Mob) and four years ago we started this project at my client. We were 12 people including the PM and Scrum Master. We were pumping out features at a good clip for about the first year. Then all of a sudden they started hiring devs left and right and we ended up with over 60 devs, split among 10 teams and productivity just ground to a halt because we're really just working on the same apps and backends, but all in our own little silos. We can't get anything into production anymore because no one has a complete picture of anything that's going on. I was even removed from the business unit that runs this project and put in a specialized unit constructed around this one feature I took ownership of. So now we have our own management structure and release schedule and we have about four standups per day, where about 80% of attendees overlap. It's just a mess.
Your PM team clearly never read The Mythical Man Month by the late Fred Brooks. You can watch a lot of precis/breakdowns of it here on UA-cam. Basic premis: You can't get a baby in one month by getting 9 women pregnant. But he goes into the details of why these teams descend into a mess.
Exactly. Many people here in the comments section are coping by saying the og video is wrong and over-hiring is not a thing. Having too many workers for a project a absolutely hurts the project and those who are disagreeing are either stupid or just coping and making themselves feel better because otherwise they'll have to contend with the reality that there are indeed too many devs in the market, thus reducing the demand of developers.
This is the problem that a lot of companies face - they have big dev teams and nothing is organized properly. The solution that is highlighted in this video is layoffs. You have a team that is too big, layoff right?
But what if we didn't? What if instead we reoganized to work on bigger and better features. What if as developers companies hired juniors to collect data to improve and refine a product. Too many developers? Hire some social media managers or marking and business assitants to help with the end product. Devs too isolated in their own bubble? Get some PMs to work with these devs to bring them back to the full picture.
Devs today are tasked with a great deal, so what if we got better about supporting the company as a hole by hiring more support roles?
Im also a consultant and can agree that bureaucracy destroys productivity
"a contractor who has to pay a cut to the Mob" is probably the best way to describe working as a consultant😆
I used to work as a consultant and I wished that project managers could be replaced by A.I, those mfs consistently made highly illogical decisions
7 years ago a professor at a notable university explained that new age business practice is firing all the people who are paid the most (and the most experienced). This is so they can hire new people for less than half the salary. He explained that it's leading to worse quality of everything. Now here we are with a growing demand for quality products. :)
I was laid off and the statement was it was because of my salary. Please take my advice and work/save to retire as early as you can. It is so nice to reach a point where it doesn't matter what happens. I wish I started saving earlier, I could have retired 10 years ago.
That’s what my last company did. The cuts were based on salary.
People wonder about why every company lays off 10%, and it's so #1 workers are scared #2 job market is weakened #3 new employees are easier to indoctrinate
I would say its closer to replacing foreign call center workers than it is programmers...
AI powered scam calls are the best use case right now - you can get a very human sounding AI powered chatbot to call people and ask for donations, while pretending to be a nonprofit etc
@@PowerWinsTopthanks bro
@@PowerWinsTop This is absolutely the risk of AI. It is very good at making things that look and sound right but is entirely fictitious. AI creates false information and scams generated at a way higher rate and far less recognizable than we've had before.
In terms of capabilities yes, but cost not really. Those workers are getting fuck all in pay, meanwhile ChatGPT's API pricing is extraordinarily high and just a single GPU to do it yourself (at reasonable speed, let alone real time) is 5 digits. Then include the fact that you also need to generate a voice for that text (TTS would get hung up on instantly) and have the hardware and infrastructure to handle all that audio data going from your servers to the phones.
For the scale these call centers work at I think there'd need to be more developments in terms of speed/price over quality
@@MaybeADragonthey don't know how little money we in third world country charge for these jobs. Lol also we have shit ton of people available too
I read somewhere that the big tech companies hoarded talent as an anti-competitive move. The layoffs are a reversal of that talent hoarding move. This creates an opportunity for start ups.
Except that we're currently in a bull market so there's a smaller than ever number of investors.
May I introduce you to Section 174 of the US tax code?
Section 174, aka order 66 for start up
@@timmygilbert4102execute order 66
I also heard a lot about this OP - get the talent we can to make sure we have what we might need but also stop our competitors getting them.
"The Brutal Truth" and then it's all just speculation and "I don't think…"
Wild speculation and weak takes
Exactly! Didn't hear a single actual logical argument.
I’m unsure the guy in the video really knows what he’s talking about.
@@gingeral253 if you know what tf you're talking about, then you're a programmer and have no time to make videos
@@monolith-zl4qt There are programmers that make videos you know. You can probably make more money by doing so.
People assume companies make smart decisions. They don't always. They saw Twitter layoffs and executives were like oh shit maybe we don't need all those people. In a couple years when all those companies are getting hacked and having failed deployments they'll be like oh shit that's why we hired all those people.
Nope, after a project is finished and has entered maintenance mode, you only need a fraction of developers to support it. The base product Twitter has already been built, and will only need a fraction of devs to support it now. So as rude as it is for Elon to fire them, Twitter really doesn't need that many devs anymore.
Then Musk had better make that point. You made the ONLY (possibly) justified reason for laying those Twitter employees off.@@Dipj01
@@Dipj01 also, as rude as it for Elon to alienate users - less users means less development and maintenance. Also as rude as it is to alienate all clients paying for the ads, less money means less need to count the money.
@@Dipj01SaaS products never are finished and just enter “maintenance mode”. That’s a pretty fundamental concept behind software as a service.
@@ThatDJMat Sure, but to implement some features, rewrite something or fix bugs, you don't need 100's or 1000's of people. So he is still right.
The thing about "coding AI" has always been funny to me because you need engineers to understand your tech stack to actually accomplish the goals of a given business need or strategy-something that necessitates human to human understanding and interpretation.
It will be an assistant for quite a long time, at least for anyone doing backend and full stack.
If a programmer is replaced with the current state of AI, it's because, either that person should have been fired regardless of AI, or the person who fired them should have been the fired one.
This video should be a tutorial on how to jump into conclusions without knowing shit. Hahahahaha
Yeah watched the whole video and don't think bro said anything profound or even accurate. I actually even reject his conclusion that you should "become more generalist to make the company valuable". How far is having surface level understanding of 10 frameworks going to get you? And if the answer is "well, don't just have a surface level understanding", then I implore you, how do you bend space and time to gain any notable experience in everything? I see so many of these tech commentary videos nowadays and at the end of all of them, I'm always left with the same question of whether or not they even work in the industry or whether or not they are grifting.
My favorite was, “no company ever needs more than 100 engineers”, but there were just so many good parts to this.
Thank you for saving my time 😂
Just like any "The Brutal Truth about xxx" out there.
@@poopymcfartbeanthat's probably a bit of hyperbole, but I agree with his take. The software companies have a hell of a lot more developers than what they produce. Having huge number of developers beyond a certain point actually hurts the project instead of helping. The classic adage of too many cooks spoil the broth still holds up.
as a math PhD student AI is not good at math in the same way a human can be good at math. it's very good at doing the kind of math you do in a calculator, but doing pure math like topography (my specialization) it just isn't all that capable of doing.
but I guess this is the difference between a mathmatician's conceptualization of what math is and what everyone else thinks is math.
topography or topology?
people call caclulation math, while math is more abstract and concise modelling
Idk its been pretty decent at explaining completed proof solutions when im stuck. But yeah it cant write a proof from scratch.
Automated proof checking is totally doable.
I disagree a bit about the generalist part - you can earn a lot more and have a lot more job security by being a specialist, it goes either way
I am also confused because did the primagen not argue for "mastering" a certain technology? 🤔
@@9s-l-s9 I think you're confusing Prime saying "don't be just a frameworker" vs. "don't be just a domain specialist". Prime is a domain specialist at Netflix.
It is true. The risk is that you become dependent on your niche. If it shrinks or collapses or goes obsolete - you might have a difficult time finding a new job.
@@Steelrat1994learn c++, it will need maintainers for decades to come, probably longer than we will work
That guy used "specialist" in the wrong way. A specialist is someone that excels at and is especially knowledgable in a specific area. But he doesnt talk about specialists, he is talking about hyper siloed roles.
Lol the disrespect of UI is crazy.
The strength of the interwebs is also it's weakness : Any goofball now has a platform speak authoritatively about matters they know nothing of.
@colincotterell3365 great UI people can actually help form strong requirements and even save companies a ton of time and money by putting up UX guardrails to minimize user error and also by reducing risk of getting lawsuits based on usability
@colincotterell3365 this has absolutely nothing to do with Linux' UI issues
my hypothesis is that corps got a bunch of free money during covid and expanded gangbuster and then once it went away they realized they had too many engineers and axed a bunch
essentially
My hypotesis is every corp wanted to create their own Uber and Doordash, and hired many web devs coders not understanding whats actually required.
Yeah Covid should go to individuals, Not companies or gov't agencies.
Govt spending was out of control, along with foreign policy, in turn the speculation that justifies tech hiring booms is less reliable. The EXTENT of the AI craze is a gamble, and confidence of meeting necessary EXTENT is quietly lacking.
@7:28 You nailed it. I'm a computational physicist and control systems software engineer: trust me when I say ChatGPT is shit at math too. I think most people just don't dive deep enough into math with ChatGTP to see the hilarious gaffes. Moreover, if I were to “accidentally” let those gaffes into my code, people would get hurt. We don't generally use things like ANNs to compute control signals in safety critical systems because we need analytical techniques to prove certain theings about program behaviour (not that you could even get insurance for it, even if you found an engineer who would sign off on it). So letting AI wirte the software in the first place is a LOOOOONG way off.
totally agree. LLMs are hilariously bad at Maths. Only people who have no idea what they are doing will be fooled by it.
The people that use AI for code probably aren't good at auditing, or even know what that is.
AI code should be put through the same filters as any other foreign code submission (like all open source code contributions).
This sadly defeats the point of using AI to begin with, that is for the majority that just wanted to do no work and still get paid.
They say that it’s good with job interview questions. I would not trust these tools with the code they spit out though.
Does it even need a deep dive? Tried asking it about energy needed to boil some water in a microwave. It gave a different answer with each click of regenerate despite seeming to use the same formulas each time.
I've only tried using it for things that I have at least some rudimentary knowledge about though. If I asked it about some legal stuff I would not be able to spot any issue ...
Was helping a friend with a pharmacodynamics problem not too long ago. Chatgpt couldn't solve a pretty easy to calculate formula ln(C_max/C_min)/delta_t. Got it completely wrong every time. I even gave it the correct answer. Different answers every time. It at least got the general idea correct. But...yeah. For more esoteric things that don't have a whole lot of training data to develop the model, I highly doubt it'll be any good. That is unless true, generalized AI actually works where it literally thinks and functions like a human with a 160 IQ, or more. Highly doubt that'll ever be a thing though.
What Thor's been saying specifically regarding Godot vs. Unity is that while "Unity - the Engine" is still kickass, "Unity - the company" has in very recent times made decisions that make them untrustworthy. Even if they retracted their utterly insane monetization plan, we can't trust them to not make stupid decisions again. Decisions that might mean they're not gonna be around in x amount of years.
+1 for encouraging kids to do highly complex manual trades. My father was an engineer and welder. He could draw stuff and then weld it together. Fuck my "IT skills".
My dude, it's not that hard, i think it's way more easier than to program
I feel like the guy in the video is a good example of fake it until you make it. Super confident, very professional video, and yet has no idea what he's talking about.
imagine having to work with that guy. insufferable
Which one?
We've all worked with engineers like this
Oh really? And you have all idea?
This is all I needed to hear.
button count proportion makes sense. There should be at most 1:2 (button:engineers) so that at no point in time a button is without an engineer. No button left behind.
I say proper redundancy requires at least 3 nodes, so I propose a 1:3 button to engineers ratio
I love sarcasm!
As someone who's been in the security industry, super specialization is very very important.
Yeah his point about being a generalist makes little sense to me. Whats the point of working on 10 different frameworks? At some point, the requirements of the company will demand specialization in different areas so your surface level understanding of things will no longer fly.
Any recommendations for someone studying to get their first network security job? I know that's a pretty open question...
@@fuckmyego sorry man. I just lucked into the job.
I've got two pieces of advice.
1) never roll your own security. Always use open sourced vetted libraries to do it.
2) by never I mean never. Unless you have a PhD or equivalent work experience
@@Peaches4Rent i really appreciate you saying 'or equivalent work experience'
Lmao talk about a fitting username too@@Peaches4Rent
The stock market works on forward projection, not current value. Therefore, hiring more devs can significantly increase your company's "value" by projecting strong forward growth and product release. At some point, you have to either deliver or make major cuts before the bubble bursts. The latter solution seems to be tech companies preferred. By doing this the company can have overvalued forward projections, and if they cut in time, they will get a second boost in evaluation for cutting costs. The workers are mostly pawns in this money-printing scheme.
The point is not that Facebook did this intentionally. They goofed. Most Valley companies goofed. They did not correctly anticipate growth decline because it never happened to them before. The valuation increases is because they actually turned around and fixed it instead of bleeding and having way too low EPS and gross margin.
Yes, it‘s a cycle, because investors are shizofrenic. When the fundamentals aren‘t there, and you don‘t have staff to fire, you hire a bunch of people, projecting future growth. If that doesn‘t work out, you now have a bunch of people to fire, projecting future growth. Rinse and repeat. This way a company, mainly in tech, can make the line go up for a long time, while just treading water on the fundamentals.
@@dv_xl they only didn't antecipate the decline if they were to dumb to understand how that growth happened, or were they counting on people working from home forever ? because the same companies dont even want their own workers working from home anymore
Wait...they dont talk about this in the video? This is the main reason for the layoffs so the fact that they didn't even cover it makes everything else theyve said suspect...
Judging by these responses, it's clear people are confused about the cause and effect of certain events. Stock goes up because layoffs does not mean layoffs occur to make stock go up. If that was the case, take it to the logical extreme, no companies would have any employees.
I absolutely 100% agree with Prime. It's interest rates, not "over-hiring"
Do some game dev myself and needed to calculate the physics of an arch projectile driving an arrow with set height and ai failed, finally found a physics lecture in UA-cam and was able to derive the correct equation, still has ways to go
What can I search to find the lecture? I'm having to work on my math a bit since I started game dev.
I tried to get it to help me write an implementation of hookes law for a vehicle suspension and it reversed the sign so that spring pressure was max at max extension rather than contraction. It took me a couple hours to figure out what was wrong as I'm pretty new at gamedev and the math around it (this is a game to learn this all on). I eventually just read the wiki page, understood it, and implemented it in less time than it originally took with GPT. Additionally, because I spent the time to learn it myself, I now have new knowledge on 3D math that I have already applied in other features. Thanks for wasting my time GPT.
Flip didn't feel like zooming in today, i respect it
Thanks for putting out this video. When Josh put that video out, I was shocked at the amount of bad information.
It doesn't take "shockingly small number of software engineers" to build and maintain the modern applications like Facebook.
I'm glad someone with a UA-cam following made a counter video
People that say that it is an easy work to make a button/form/etc. - haven't really made one, right? Making a fully-featured set of controls that needed to make an excellent form - it is hard.
And i'm not even talking about the fact, that making uikit is not what most frontend devs do daily.
Localization, accessibility, internal tools...
Nothing like mass layoffs at the top companies while I'm doing my job search to be able to move !! So fun
We were in this situation at the end of 2022 through 2023 and I got a new job July last year. If you're good at what you do don't worry.
I graduated in the height of the financial crisis, got a job in finance, and has since 5x my income. Who you are can make a big difference.
@@jimbeam9504just because you managed to land a job doesn't mean a job crisis doesn't exist. People not landing jobs doesn't mean they're not "good". Survivorship bias at play there, mate.
31:15 Hard agree with Prime. Honestly, even from the perspective of a Gamedev, this is still shortsighted. Most games won't need 100 people, but the truth is, there isn't really a limit to the feature-set of a game.
Even taking a simple game like Minecraft, you have:
-World generation
-Interactive blocks
-Movement with / without items, while / while not riding, all of the movement stuff
-Crafting, enchantments, repair, etc
-Fighting (armor, potions, effects linked to enchantements)
-NPCs
-Shaders & VFX
-SFX
-Optimization
-Network code (LAN & server)
Oh, did we forget the dimensions?
And the command blocks?
And the events?
What about the tutorial?
Of course the recipes.
Wait, there's a non-java version? And console versions? And even a phone version?
None of these tasks are hard or very lengthy, but they can be worked on by different people at the same time, which is the most basic requirement for hiring employees.
So turns out, if you want to make a Minecraft clone really quickly, you could have maybe 20 or so engineers.
No one would hire so many people to deal with so few simple tasks, but it's possible.
Now how would you even manage to convince anyone that a project can't be split into more than 100 parts?
he is pretty naive because he is a game developer, I was a game dev then went to web for a better life, but it is incredible the difference between people (humble people) compared to the no soul that exists in the game developer industry. The whole mentality is that everyone dont know sht since they are not doing game dev.
Based on your channel, you are the farthest thing possible from calling your self a game dev..... script kiddie is a more apt description of your talents.
The last part is kind of true ... :))
28:02 - What... What just happened??
On feature delivery times: localization itself is a feature that greatly impacts the delivery time of next features. Last project i worked on implemented it and suddenly we had a new step before release - send the feature to the localization team. It added between 2 and 5 days depending on what we were delivering
Facts.
6:35 I think his point was more about the creativity argument, and I fully agree with it.
People love to think of creativity as something unique that AI could never replicate, but this is severely mistaken.
Precision is indeed the hardest problem that likely won't ever be fully solved, as it has infinite definition, but what we call creativity is really only the sum of our experiences mixed with randomness. AI is currently not great at being creative, but it is a finite problem. Maximum creativity is random noise, and what we consider as being creative is within bounds. There's the randomness you want (creativity) and the randomness you don't want (bad creativity), being creative is just having an sense and intuition for a subset of the valid creative space.
Twitter is not a billion dollar idea. It is a -40 billion dollar idea 😂😂
it clearly wasn't lol. it was so silly of elon to hand twitter a giant bag of money.
big if true
Twitter was a proven 40 billion idea, X is so far a -40 billion idea. lol
@@melski9205 its said that twitter was rated at higher worth than it actually was, i think it was aroun 20-25 bil
Elon hate is media brainwashing. Dude is literally the most successful entrepreneur of all time. Nobody else has started an American auto company in recent history. Few have created space companies (and SpaceX is on top). Starlink is bringing the Internet to everybody, everywhere. Elon buying Twitter allows free speech on a platform that previously was caught suppressing "unpopular" views in collaboration with CIA (see twitter files). Everything I've stated is known facts, and all you've got is "lol 40 billion big number"
Chat Gippity is a tool, like anything else. Yes, you need 100% working code (which AI rarely provides), but it doesn't matter. The point of it (and CoPilot) is to write some boilerplate code, as a "springboard" for you, in order to reduce dev time. Also, Chat Gippity is great for getting answers to high-level concepts. It's not just good for writing code, but I'd say it's almost spot on as a learning tool to give you context.
DISCLAIMER: Yes, don't trust it. AI hallucinates. Always vet and test its answers. Just don't throw the proverbial baby out with the bath water
Twitter is definitely not faster than ever. Like it'll routinely just give up attempting to stream a video about 5-10 seconds in.
Brett Victor gave a brilliant presentation called "the future of programming". In it he talks about coding in goals and constraints, not instructions the computer should figure those things out. This concept is over 50 years old now, maybe AI is how it will be unlocked. So not replacing programmers but changing the way they work
51:00 a previous company for who I worked had a c level employee 'reply all' everytime he needed his assistance betty to post on social media. he was also the reason why the implementation of 'password change every year' because he kept clicking on phishing mails....
Thought experiment...let's say we didn't have any type of sorting algorithms yet, but we do have AI. Could AI come up with all of them? If yes, why hasn't it come up with any new ones yet?
Slack notification got me good
The worst of all possible jump scares
Well, its a bit more than a complex approximation function. First, its a recursively self calibrating function.
Then after a while you realize that its not only approximation, its actually a deep linear regression, that ultimately will linearize an input set of non-linear values into recognizable output of values of a well defined manifold.
I was half way through the wheel of time books and i asked the AI about my progress because i had some questions, and it was telling me all kinds of crazy shit that never came true.
i got laid off jun 5. the silly part is it had nothing to with the trends in the industry. our ceo was caught embezzling money and we lost all of our contracts. now im dealing with the state of the rest of the industry. life is funny that way. pray for me
I like the "wtf this is dumb imma rewrite it?" And then trying. I learn a lot. It's fun.
Half the time you do end up with something very similar to the original. :)
I absolutely think *some* programmers will lose their jobs to AI. Especially in non-tech spaces. IT/Engineering HR competence in non-tech environments is not great. I mean if you're on COBOL and IMS you're probably safe from AI, but if you're specifically doing reasonably modern tech at a place run by AI enthusiasts there's probably a non-zero risk. However, that's pretty niche and even then I doubt the threat is substantial.
I have a pdf copy of a paper written by Nathan Papke that talks about AI with great detail. It even includes some C for loops that mimic rudimentary neurons. I didn't know it was so long running until I read that.
A lot of tech companies have workers who legit do nothing.
(If you don't think this, you are one of the people who does nothing)
The higher-ups are slow to realize that their company has bloated, seems to come in waves.
The thing is it will not flat out replace you. It will make your job take less time. Which probably means less programmers will be needed. Hopefully this is offset by all the jobs AI created for when building AI-stuff.
But it's best for the very tedious stuff. Like populating a large json or struct.
Like in many earlier instances of productivity increases, i'm not even certain the total amount of programmers will necessarily reduce, atleast not for a while. We'll just increase the scope of the things that we're trying to do instead.
Wouldn't that just mean more tasks? There's always things to be done
It boils down to a single reason: money. Layoffs occur because companies can no longer write off 100% of software developers' salaries against taxes. Have a look at Section 174 congress act.
Zelda BOTW and TOTK spent like 300h in each and I wouldn't say there are no bugs, but so little that it's unbelievable
7:00 there's a whole lot of people making decisions in business that are trying to get by with "close enough that others don't catch on," and the truth is that it catches up with you, instead of people catching onto you
6:55 I intentionally put that top-level comment a few seconds after the point I was responding to to illustrate what "close enough" is like
On whether AI can solve new problems: we're used to thinking about AI as a tool to make creative works from prompts, but a lot of the power will be coupling it with skilled humans in a conversation. Instead of just Jeff Dean and Sanjay Ghemawat pair programming together, imagine them programming with Gemini. Or imagine Donald Knuth paired with an AI. Or check out Terence Tao's explorations with using ChatGPT in proving theorems.
Likewise, while generated code is currently a nightmare, I suspect people may underestimate the savings we'll get from intermediate and advanced engineers using AI in the design (rather than coding) phase.
Wow somebody doesn't understand how AI works... it isn't thinking... it is a probability machine that based its probability on a bunch of stuff that people wrote. The only thing that is being underestimated here is how dense you are.
@@thomgizziz are you grumpy?
@@thomgizziz are you under the impression that the human brain is not a "probability machine" (as you put it)?
I have spent time pair programming with Gemini. For very small snipets in simple languages like Python, it is okay at best. For anything else, it is terrible; Calling methods that don't exist, using types that aren't defined, making copies of heap allocations where they are not needed, exponential time solutions to linear time problems, testing things that don't need tested, not testing things that do, using third party libraries that don't exist. Design isn't much better. AI works great for 'best fit' type of problems. Programming and design don't generally fit into this category.
@@ants_are_everywhere What makes you think it is? Yes probabilities are taken into account but you don't make sentences by choosing the next word you are going to say by taking a list of probable words with percentages attached to them and then taking a random number and choosing that word because the number told you to and it might have been the least probable and therefore nonsense.
Stop it with the bull.
The main reason for the Tech layoffs is the increasing cost of capital due to rising interest rates. For the longest time, capital was free and making profits could be delayed until the future. Now capital costs money every day, increasing companies incentive to shift focus from stock market value to actual earnings.
Do you mind explaining a bit further. I am still confused..Thank You 👍
18 minutes = 1 hour in prime time
AI developer and autonomous aerobotics engineer here. Def not in an nft FOMO type deal due to actual physical MVPs in our world. There will still be a need for software devs, engineers and programmers just like we need car mechanics for our cars.
I totally get with the 'NFT era of AI' take. GPUs getting pricier by the day, companies funded by investor hype and not much else, it's echoing that era so perfectly you'd think it's parody
A few years ago everything cloud was uber-hyped, today it’s AI. I wonder what’s the next thing.
@@andrzejostrowski5579Quantum ?
@@andrzejostrowski5579 i mean cloud computing was very much a game changer, you will be hard pressed to find a startup today that doesn't run in the cloud. NFTs on the other hand...
I've long held the belief that corporations tend to over-hire during economic booms, possibly with the intention of identifying and retaining the best employees while also providing a buffer for future needs. However, when economic conditions shift, they may seize the opportunity to part ways with underperforming staff under the guise of economic necessity.
around 23:34 your red bar (video progress) catches up to Primeagen's.
To answer the question about why you still need a software dev for things that already exists. It's because a normal person doesn't think like a developer. A normal person can't even put in the right prompts to get the desired output. They need to think about all the small things, how it flows, common features, and missing features.
AI can't really code that well right now, but even if it could, people can't think like a programmer. It's like having a team full of bright people, but they have an ignorant lead who doesn't know how to use their skills.
AI for developer is rubber ducky on steroids.
There's a problem of personal interest
No manager says "hey boss, I have too many engineers, you can take one of my best and move her to another team that needs talent"
No director ever says "Hey boss, I have too many teams in my org, you can take one team and disperse them to other orgs"
Managers always want more money more engineers more resources, and inevitably they will overhire if the company lets them
42:00 is probably the best advice you can give new or Jr programmers from my experience. you can get really good at something and that's great, but being flexible and having the ability to understand what people are saying across the board saved my ass more than anything else.
"Be a good programmer, and also a generalist" means "have decades of experience"
Thanks for the react, Prime! Loved hearing your opinions, especially coming from somebody so experienced. Cheers to many more years in tech! 🎉
7:16 Physics major here, AI is horrible in math. Not only can it not do any advanced stuff like Lagrangian mechanics for even the easiest problems, but the other day it couldn't generate a system consisting of 3 equations with 3 unknowns (essentially a 3 by 3 Ax=b problem) which I asked for my 9th grader student.
Tech companies are laying people off for a few reasons:
1. Investors love it as it increases short-term revenue, so tech companies will continue to do it while they're being rewarded.
2. Employees are costing more due to wage inflation.
3. Related to #1 but Interest rates are higher which is hitting the bottom line meaning laying off is an easy way to show you're focused on maintaining your margins.
4. None of the big players are investing in creating new traditional apps or services, they're focusing on growing what they already have. You don't need as many engineers for maintenance as you do for creating something new.
5. No one is creating traditional apps + services because investment has swung to AI, no one wants to invest in the "old" stuff anymore. Most apps that involve AI don't need that many engineers to make as they're all utilizing existing compute architecture (which handles large-scale distributed inference and has existed in AWS, Azure, GCP, etc for a while) and most AI apps are just relatively thin wrappers around calls to LLMs.
37:17 so apps with infinite scroll feature with buttons added each time need infinity software engineers?
You're onto something here
Very interesting takes about generalizing and specializing, I've pretty much always heard that it's better to specialize in something because you'll be incredible at that one thing and in the same way be hard to replace
There is such a thing as specialization death. If you are not generalized enough if the market changes and what you are specialized in is no longer needed you are in a dead in. Both over qualified and under skilled for available jobs.
@@thomasmatthews7388Of course but that just means you need to be able to adapt to the market demand. I don't think the solution to this is to be a "generalist" because you will only ever have such limited understanding of the things you work on. Unless the company isnt evolving, eventually the requirements of the company will require you to start specializing in certain aspects and the generalist approach won't cut it. Specialize, but don't put all your eggs in one basket is the better approach.
@@thomasmatthews7388 But that implies that you've already hit the threshold of specialization, and then you start diversifying. Not that you're the guy who knows a little bit of React, a little of Django, a little of SQL, a little of devops, and you're trying to be a one-man band pretending that you can ship at the speed of 5 people.
The answer is T shaped specialization and generalization.
T shaped knowledge is still the way to go.
58:13 Flip, my dude, you're dropping the ball, brother! 😂
More than 1 hour, but a valuable content!!!
The layoffs are more economic than AI or other phenom.
Learning fast is a huge benefit everyone that is a beast that I knew they have these particular skills, they're learning stupid fast new things!!
This was a good video. Always love the balanced takes Prime.
He doesn't have the full picture on the layoff, which he says are bad for the company, but there's an investor hype cycle these tech companies leverage to do enormous buybacks right after layoffs.
19:54 You just said that you can't compare an house with a skyscraper, and yet you're comparing twitter with AAA games
the man overcooked
Questions I have about this: (that I will be looking into anyway but anyone please let me know if you have the answer.)
1. Are these 1000's of tech layoffs all definitely developers/engineers?
2. What does the job market look like after this, are people fighting over a small job pool or will we see a lot of new start-ups trying to disrupt big-tech?
3. Is freelancing a better path for the industry as product requirements seem to vary too much to hire permanent staff?
I recently said something stupid like "CEOs do nothing. They spend 80% of their day in meetings. I spent a lot of time in meetings too, and nobody's paying me a million dollars a year." I got a good reply: "the fact you don't see the value in what a CEO is doing is the reason you're not being paid a million dollars a year".
I'm not an expert when it comes to the economy or Wall Street, so is there anything that prevents a company from "burying" their real revenue/profit margins with the cost of hiring a bunch of people for the sake of "preparing for a future project"?
I was under the assumption that home investors/professional investors could be tricked into thinking a company is more valuable with employees, not just because of the increase in headcount, But because the cost of hiring those employees might be used to fudge other numbers.
I'm pretty sure that I'm wrong but I just want some help understanding why.
It's hard to hide the lack of growth. When companies are burning cash ( e.g. Amazon in its first decade or so ) they are given quite a bit of leeway if they are demonstrating significant growth. The concept being that funding growth is costly ( new regions, new product offerings, marketing / customer acquisition costs, etc ) and if the growth is there costs are discounted. Although if costs YoY are outpacing the growth metric, then the knives come out.
So no, you generally can't hide a fundamentally bad business by increasing costs uncorrelated to a growth metric that the street accepts. Now this doesn't mean that some bespoke growth metric isn't B.S. , but the street's analyst community usually smells the 💩 pretty quickly.
You're my favorite dev UA-camr man. Based takes across the board. Inspiring devs to grow technically with good character. You go gurl 🎉
Always nice to watch Primes reactions, because you're essentially getting 2 videos, 2 opinions in one video
They can afford to fire X% of their workforce even if they have temporary losses, make whoever stay work more hours or more intensively and, in the next decade, tech salary will decrease for everyone, as there will be a bigger mass of unemployed developers desperate to work for much less than what they used to.
Every industry has done that over the past decades. It's just for greedy profits.
and yet there is a programmer shortage :/
I think skill comes into it, lots of new frontend devs, not as many backend with the experience companies need.
the pay will likely be polar opposites for a while
there is always a shortage of everything if you ask companies. if i may translate that for you, it means: "we dont wanna pay our workers, please overflow the market with supply" @@Mel-mu8ox
Didn't even know TechLead said that. I watched a few videos like 4-5 years ago, found that he centers everything about him having supposedly been at Google and now a millionaire, found him annoying mainly due to his personality - and clicked "do not recommend channel" every time I've seen him again.
Good to see that I had an accurate reading 😂
how many of the layoffs were actually devs?
Right now it’s mostly devs not being hired rather than devs being layed off. Thats because companies are anticipating that AI will be much more intelligent in the near future, so there’s no need to hire right now. Eventually it will replace devs in a few years as it gets more intelligent exponentially. I know he doesn’t agree but it’ll happen just wait and see
1:11:15 While both Nokia and Ericsson left the consumer electronics market, they are still the second and third biggest vendors of telecom network infrastructure at around 13-15% market share each. The pair are building most of the 5G infastructure in countries like the US that banned Huawei, and Nokia Bell Labs is even building a 4G network for NASA on the moon.
Spoiler alert: it's just a result of overhiring before and during covid
lets see if i say this
@@ThePrimeTimeagenYou are watching with us? :,]
I think we should stop considering head counts in tech space. It just doesn't make sense. One person created Instagram but his motivation was on a different level and now you need 100 employees to mange the same app because not all devs have the same level of motivation. They just work 9-5. And there is nothing wrong with that.
AI will not replace you in your job. Somebody who has your skills, talents, abilities and knows how to use AI, Will replace you in your job.
That's what I've been saying. No one seems to want to counter argue that. They are in denial.
And AI may also soon replace your job. These dummies who use today's AI as reference and assume nothing will progress non-linearly are batshit, utterly.
Thank you for the BG3 callout, its legit a piece of interactive art
I belive that a major reason for layoffs in Tech is simply because they over hired people during the tech boom a few years ago.
And now there are so many useless positions occupied by people with little or no reason to be in the company to begin with.
DEI "officers" at my workplace for example, where all fired because they never did anything that would profit the company.
People who go to meetings just to be in meetings also got fired, they never actually provided anything of value to the company.
This is also why Elon fired 80% of Twitter when he bought it, and it still runs just fine, because the majority of people working there weren't necessary for the company to work.
Did they rehire new DEI officers? Hopefully they did but if not, thats very concerning.
DEI in this industry is the worst I’ve seen/experienced in my work life so far. I haven’t worked in too many industries but its a major problem.
@@xmissyangelzx
They have not rehired the DEI officers. But as a response we have apperently lost some funding from the orginazations that demanded them hired in the first place.
But overall its still a money saving decision. The DEI officers did nothing but hinder the hire of great employees because they didnt fit the quota. Being Diverse (Not white, straight and male) is what they strive to hire. This is terrible for buissnes in general, because most people in gaming are straight white guys who do nothing but game all day, and to refuse to hire those with the passion to work for games or codes because they are not meeting the DEI requirements are a loss to the company.
The NGO's who wanted us to hire these officers in the first place has made sure we will never be in the Game Awards, or at the very least threatened to.
@@paragonaesir1957 I understand that and it is a fact that most developers are white men/asian men (unsure about sexuality as thats harder to track). But they shouldn’t have too much trouble finding people that arent white /asian men for other roles in a gaming company.
Im not in the game industry, Im just a mobile developer. But I will say DEI is really important for the gaming industry because gamers come in a variety of people that aren’t just white men. When you have a diverse team, it brings new perspectives which lead to new ideas, as well as accurately represent the global audience that makes up gamers.
I understand the company you work for doesn’t value it but thought to say this.
DEI isn’t just about hiring diverse candidates, it’s also about training people appropriately to know how to treat others different to them.
Discrimination is such a personal issue of mine when it comes to my work, Im at the point where Im gonna leave the industry if I can’t land a job in FAANG because these are only set of companies I see that actually put in effort into DEI, particularly on the engineering end of things.
I hate the Data structure and algorithms stuff i have to learn but I see no point in me continuing this job, if I have to feel like Im fighting with my coworkers all the time, just to be taken seriously at my job or even just simply do it lol.
When I mean its such an issue, I notice these issues from interviews I do as well as every place I’ve worked in this job at so far.
Im very much an outlier in this industry. But its been more of a handicap than anything.
In comparison to my white male work colleagues they don’t have such hard time with how people treat them, despite us all being as competent.
Also in general, most women Ive worked with have mentioned instances of not being treated well, for unjust reasons. Theres a-lot of sexual harassment, bullying, racism, sexism, (i dont see much ableism as im certain loads of devs are autistic😂) but theres just alot of bad behaviour people do intentionally/unintentionally that people get away with because of the lack of effort put into DEI. Theres also statistics that most women leave the industry.
Anyway I don’t mean to ramble. Im not into the idea of hiring someone to fit a quota but most places won’t hire someone that cant do their job as businesses need money lol.
But I just thought to put it out there. I can’t speak for how your workplace handled DEI as Im not there but thought to just give you another perspective.
So, ~20 years ago when I people used to ask language trivia in interviews to show mastery. When I did my most recent round of interviews at FAANG companies, they didn't even care if I was familiar at all with the domain or language. They just wanted to test for general coding ability and problem solving. So right off the bat this guy is not even a little wrong, he's Uncle Bob levels of wrong.
He then says "social media sites used to be developed by 1 person"... I actually think this guy is a master troll.
Wait till he gets to the part where he says if you have over 100 software engineers "you're done". As if 100 is optimal, but 101 is not. He is probably a troll.
AI doesn't replace programming jobs; employers do. They could be socially responsible and retrain people instead of firing people and bring on new people who aren't familiar with the company and require costly on-boarding.
My concern with AI tools, is the same I had with outsourcing jobs. The advanced jobs aren't going away, but advanced programmers all start as entry level programmers who get better over time. If you need to have advanced skills in order to get a job, a lot people aren't going to get the real world experience necessary to get better and a lot of people won't even get into the industry at all which will lead to longterm problems finding advanced programmers.
AI has shown me the arrogance in the programming community. 5 years ago people were saying artists were gonna be the last to be replaced by AI because their work requires creativity.
I know companies, not yet public, that use VC funds and hiring people like crazy. And they do all this show to portray a big company. So, Josh is right! Bang on target
7:15 I decided to test this so I paused the video and threw it a few algebra 1 questions and it surprisingly got them all right. (I haven't used chatgpt in a while but I remember being terrible at math) so after that I thought to myself about how chatgpt was better with polynomials than me, unpaused the video, and was told that I sucked at math lol
I put in it a high school problem which had calculus 1 and differential equations.It could solve the differential equation because it was classic and there were online for similar types. But when it came thinking outside the box it failed. You just have to throw it problems made by a mathematician in your region(Each region has different philosophies of math problems) and will fail miserably. Don't just use USA derived math questions.
@nikos4677 Yeah, it more or less just memorizes the answers and looks for a few basic patterns. It doesn't really understand what it's trying to solve or what for. But it still did surprisingly well, in my opinion.
100 programmers is not a stupid take. Look at natural human social group sizes, it matches a lot of things, you have 4 good friends, 10 okay friends, 25 friends of friends you can hang out with, 40 people with whom you might have to interact this week, 100 people you have in your phonebook for this month and 200 to 400 people is all you know. These are all natural unit sizes for humans used in the military, company organization and school organization. Wherever you go if a school has more than 50 people that's two classes of 25 people and all of them will be in friends groups of 3 to 9. Those are just platoons and squads or departments and task groups.
If your company has 100 programmers and 4 to 6 departments with 16 different task teams, bro that's the most you can go with one CEO naturally. Everything above that the CEO relegates remembering names to managers. Up to 100 he can remember all the names actually.
Going deep into the rabbit hole here, Prime. War is costly, and cost cuts always come from where the rope is thinner: the working class. This has been so for millenia, and if someone is interested: "The Theory and Practice of Oligarchical Collectivism", the fictional Book wrote by Emmanuel Goldstein, the public enemy #1 of the Party in Goerge Orwell's "1984", explains the role wars play in contemporary capitalism quite well.
A man with a macroeconomics class under his belt could summarize the real reason with two words: interest rates.